This article revives the discussion over measurements of validity in criterion referenced (CR) tests. It presents how the principles of Classical Testing Theory (CTT), normally associated with norm-referenced tests, were applied to the Business English achievement tests at the University of Economics, Prague, Czech Republic. Firstly, measures of validity in criterion-referenced tests, test purpose, and test specifications are discussed. Next, a 10-item vocabulary gap fill subtest is subjected to a detailed analysis through the use of facility and discrimination indices. Key and distractor analyses of each item are then performed. The insights gained from such analyses are examined in relation to the cyclical test design process of constant review of items so that a high level of standardization is achieved. This paper thus provides teachers with simple tools to build valid language gap fill tests which reflect the criteria of accurate and equitable testing.

No abstract

No abstract